project-dev — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited project-dev (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
| Location | Purpose |
|---|---|
plans/features/*.md | Active plans (unstarted or in-progress) |
plans/features/pending/*.md | Partially implemented (gaps documented in parent dir) |
plans/features/gaps-*.md | Consolidated remaining work from related plans |
Plans typically contain:
a. Read the specific section carefully b. Note the listed files, test cases, and acceptance criteria c. Implement changes d. Run specified tests e. Verify acceptance criteria
When breaking a plan into tasks:
A task is straightforward (→ @engineer agent) when:
A task is complex (→ @architect agent) when:
The /fleet command automatically decomposes work into subtasks and dispatches to custom agents based on their descriptions. When a plan has independent tasks:
/fleet to execute — it will route simple tasks to engineer and complextasks to architect based on the agent descriptions
@engineer fix the dd_version bug or@architect design the new pipeline
Rate each task 0-3 on each dimension:
| Dimension | 0 (Simple) | 1 (Moderate) | 2 (Complex) | 3 (Deep) |
|---|---|---|---|---|
| Files | 1 file | 2-3 files | 4-6 files | 7+ files |
| Ambiguity | Fully specified | Minor gaps | Design choices | Research needed |
| Dependencies | None | 1 internal | 2-3 internal | External + internal |
| Testing | Add to existing | New test file | New test patterns | Integration tests |
| Risk | Local change | Module-scoped | Cross-module | Breaking change |
Total ≤ 5 → @engineer (Sonnet 4.6 — fast, precise) Total ≥ 6 → @architect (Opus 4.6 — researches then builds)
plans/features/<name>.md → Active (implement from here)
plans/features/pending/<name>.md → Reference material (gaps documented)
DELETE the plan file → Fully implemented (code is the doc)Every completed plan must update affected documentation:
| Target | When to Update |
|---|---|
AGENTS.md | New CLI commands, MCP tools, config, workflows |
README.md | User-facing features, installation changes |
plans/README.md | Plan status changes |
.claude/skills/ | New reusable workflows |
docs/ | Mature architecture documentation |
| Prompt templates | New or changed LLM prompts |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.